ThinkRun: Screen Recorder for AI Agents
Screen recorder for AI coding agents: turn a browser session into structured context your agent can act on.
As of June 2026, ThinkRun: Screen Recorder for AI Agents has — users in the Developer Tools category.
Usersno change0%
—
—
Ratingno change0%
—
— reviews
Reviewsno change0%
—
Version
0.2.7
Manifest V3
History
1 snapshotsTracking since Jun 17, 2026.
Not enough history yet for this metric — the chart fills in as we collect more snapshots.
View as table
| Date | Users | Rating | Reviews | Version |
|---|---|---|---|---|
| Jun 17, 2026 | — | — | — | 0.2.7 |
| Now | — | — | — | 0.2.7 |
Permissions & access
- Permissions
- tabCapturescriptingstorageactiveTabtabstabGroupsoffscreennativeMessagingdownloadsnotificationsalarmssidePaneldebugger
- Host access
- <all_urls>
Screenshots
About
Stop screenshotting bugs and typing paragraphs to explain them. ThinkRun is a screen recorder built for AI coding agents. Record any browser session — or let your agent record what it built — and ThinkRun turns it into structured context your agent can actually act on: every click, the console, network requests, your voice narration, and screenshots, all time-synced. Not just a screenshot. Not just a video. WHAT IT DOES • Screen + voice recording: capture any tab with audio and microphone, right from Chrome • Structured capture: clicks, console logs, network requests, and screenshots, all timestamped • AI-ready bug reports: hand the recording to your agent (or a teammate) as a structured, replayable report • Real browser control: let Claude Code, Cursor, or any MCP client drive the Chrome you already use — your tabs, your cookies, your logged-in sessions • Shareable links: generate a link that's also available as LLM-ready Markdown or JSON to paste straight into Claude HOW IT WORKS Install the extension and open the side panel. Hit record to capture a session, or connect your AI tools so the agent can navigate, click, and screenshot on its own. ThinkRun packages what happened — video, transcript, actions, console, and network — into context an agent can parse in one pass, instead of you re-explaining it. WORKS WITH • Claude Code, Cursor, and any MCP-compatible coding agent (npx @thinkrun/mcp) • The thinkrun CLI for local, scriptable browser control • Local mode: your agent controls your real Chrome with no cloud browser to provision USE IT FOR • Showing your coding agent a frontend bug instead of describing it • Recording a workflow once so an agent can repeat it • Sharing a session with a teammate as a replayable link WHY IT'S DIFFERENT Screenshots lose structure — an agent can't tell which element is broken from pixels alone. ThinkRun keeps the structure, so your agent gets sight that actually understands the page. PRIVACY • Local mode keeps page content on your machine — nothing is sent externally • Recordings save to your Downloads folder and leave your device only when you choose to upload or share • The extension bundles no third-party ad trackers or browsing-history collection; full data practices are in our Privacy Policy Learn more at thinkrun.ai
Technical
- Version
- 0.2.7
- Manifest
- V3
- Size
- 432KiB
- Min Chrome
- 88
- Languages
- 1
- Featured
- No
Metadata
- ID
- affbcanbhbnolapiokgjnkfjccdmefmd
- Developer ID
- ua7c222bf6c73801a99b0f427925b6111
- Developer Email
- [email protected]
- Created
- Jun 16, 2026
- Last Updated (Store)
- Jun 16, 2026
- Last Scraped
- Jun 17, 2026
- Website
- —
- Support URL
- https://thinkrun.ai
- Privacy Policy
- https://thinkrun.ai/privacy
Data sourced from the Chrome Web Store · last verified Jun 17, 2026.